Lightning Talk: Leveraging Kubeflow for Next-Generation AI Healthcare Solutions
Lightningat Kubeflow Summit Europe 2024 · on Tuesday, 2024-03-19· at 17:05· Pavilion 7 | Level 7.3 | Room S02·
summary
Roche describes a Kubeflow-based mlops platform that actually runs drug discovery predictors at scale using Kubeflow, Knative, Helm and Argo CD, plus an inner-source Kubeflow distribution and an mlops SDK and CLI. The platform serves over 100 predictor endpoints, retrain hundreds to thousands models daily, hits 100% success rate for inference at 1M predictions/minute with 2–25 ms latency per molecule, and uses autoscaling patterns to meet peak demand.
AI-written summary · about
chapters (6)
- 0:00Introduction and background
- 2:58Platform requirements
- 5:40Kubeflow as a solution
- 7:48Performance and architecture
- 10:38MLOps skateboard model
- 12:31Collaboration and future
related talks
27mKubeflow for enabling AI Powered Drug Discovery and development...
30mKubeflow, Your ML Tools on Kubernetes
14mKubeflow: Where ML Meets Kubernetes
35mBuilding a Medical AI with Kubernetes and Kubeflow
31mKubeflow - Data Science on Steroids
51mMachine Learning using Kubeflow and Kubernetes
31mKubeflow 2026 State of the Union: Orchestrating the Full ML Lifecycle
44mEngaging the Kubeflow Community: Building an Enterprise-Ready AI/ML Platform